雪花：虽然使用结果缓存，但相同的查询需要更多时间

Question

I'm new to snowflake, Using JdbcTemplate for executing queries.

 BasicDataSource dataSource = new BasicDataSource();
 dataSource.setUrl("jdbcUrl");
 dataSource.setDriverClassName(SNOWFLAKE_DRIVER_NAME);
 dataSource.addConnectionProperty(CLIENT_SESSION_KEEP_ALIVE, "true");

 JdbcTemplate jdbcTemplate = new JdbcTemplate(dataSource);

Facing below issue on Snowflake:

If I execute the same query multiple times, and gap between the same two queries <= 150 seconds , then it takes around 300ms . If the gap between the same two queries > 150 seconds , then it takes around 1200ms . These timings are from java code. Although all queries are using a Result cache because I'm executing the same queries multiple times.

Anyone have an idea, why is this latency is occurring if the gap between the same two queries is over 150 seconds?

Answer 1

At Snowflake side, the AWS load balancer has 150 seconds timeout configured on it.

Network sockets are getting closed after 150 seconds of inactivity. This is why all subsequent communication attempts have to re-initialize the connection. That's why latency is occurring